NZGDB Newsletter #6, 20th November, 2007

 

Since the last newsletter there have been a large number of changes to NZGDB.   Of course most are pretty minor, but the major changes have been: -

·         Some reports and Printer-friendly page options have been provided. 

·         Facilities for paying subscriptions have been developed, in preparation for next year when subscriptions will be required to access information other than your own.

·         Minor changes include the separation of softlinks into GDB Links and Scrapbook that was discussed last month, minor formatting changes, and error correction. 

·         Another design question.  Is it possible to “score” records, so that when there are duplicates the “best” record is presented first?

Reports and Printer-friendly page options.

While NZGDB has been good at presenting information on line, it has not been so good at providing information as reports or printouts.   Printing the individual page (GDB2) can be particularly frustrating.  If there are more than five facts and notes, only the first five are displayed, with a “Next” command to show you the next five.  Only the first 96 characters of each note are displayed: to see the full text you have to click the note.  If one of the scrapbook documents has been designated as a primary document, then only a small part of it is displayed and you have to use a scroll bar to see it all.   You simply can’t display all the information about the individual at once, and of course if you print the page (using the browser’s print button) you print only what is displayed.  

 

A start has been made on fixing this problem.  The comparison page (GDB4) and the display-document page now have a [Printer Friendly] button. Click this and a version of the page is displayed that lacks the usual graphics and links that appear at the top, left, and bottom, but more importantly on the comparison page the complete comparison report is shown, and on the display-document page the whole document is shown, not just a scrollable section of it. For example, click this to see an example of a comparison page, and click the [printer friendly] button to see the alternative layout.  Click this to see an example of a document display, and compare the [printer friendly] version.

 

Even greater changes have been made with the Individual Page (GDB2).   Here the equivalent button has been labeled [Reports] because it is more than merely a printer-friendly version of the standard page.  Clicking [Reports] produces a page where: -

1.                   All facts are displayed, not just five at a time.   If the facts have attached detail records, such as source information, then this is displayed also.

2.                   Notes are all displayed in full (not merely the first 96 characters).   Notes are displayed separately from other facts.

3.                   If one of the scrapbook documents has been designated as a “Primary document”, then this is displayed in full, not partially displayed within a scrollable section.

4.                   There are also buttons for further reports, [Pedigree] and [Outline Descendant].  NZGDB is intended to eventually provide a complete genealogy system, so at least the basic reports that you’d expect need to be available.  The first two of these have now been provided, at least in prototype.  Click [Pedigree] to produce a report of all ancestors of the current person, and [Outline Descendant] to print all descendents of the current person, or at least all that you are allowed to see.  Of course the same privacy rules have to apply in reports as apply in normal web pages.

 

This is intended to be just the start. The format of these reports needs to be improved, several more reports are needed, and some of the reports should provide options.   I’d welcome some feedback.  What do you want?  What are your priorities? 

Paying Subscriptions.

In January the site will start requiring users to pay a subscription ($25 per year) in order to look at records other than their own.   There will however be a credit for putting records on to the site (for example 2000 people gives you a year’s credit), so contributors won’t have to pay a subscription for some time: indeed, some contributors will have years of free access. 

 

Code to handle subscriptions and manage membership period has been developed.  Payment may be made by cheque, Internet banking, or by credit card.   Note that NZGDB will not hold, or even transmit, any credit card information: all credit card transactions are handled by a secure, external gateway (Payment Express) that has been approved by the banking system.  NZGDB is merely told the transaction ID, and whether the transaction has been successful or not.

 

Currently the only restriction is in the search function.   Searching will show you a list of names as at present, but if you click on a record other than your own and you do not have a current subscription then (after 1st January) you will be prompted to pay.   However you can still follow links from your own records, including links to other people’s records.

Other changes.

Click here to see the full list of changes that have been made in this period.  A few are worth commenting on:-

294              Increase the search limit.  The search will now return up to 1000 records.   They are still returned all at once, rather than 100 at a time.

350            Softlinks have been separated into “Scrapbook” and “GDB Links”, mainly because I felt that this would be easier to understand.  Also, on the individual page (page GDB2) children are now ordered by year of birth, with those whose year-of-birth is unknown appearing first.

374, 377, 380, 380.   Several improvements were made to the “Manage your GDB Information” page.   The most significant was that it is now possible to see the actual list of records that are duplicates of yours.

Ranking Search Results

Often a search will return several records that are clearly duplicates.  For example, search for Hannah OLD and currently NZGDB will return 16 records.  Eleven of these are clearly the same person, who was born in 1860 and died in 1939.

Search results are returned in order of “namedate”, the standard format of {FAMILY NAME, given names (birth year, death year)} that is widely used within NZGDB, so of course the duplicate records are grouped together.  It would be great if they were listed within this group in order of “quality”, so that the first of the ten OLD, Hannah(1860-1939) records is the best record, while records that are merely copies of old information derived from an emailed GED file appeared later.  But how do we tell?  In fact, what does “quality” mean?  We can only look at each of the records, and judge which to believe in cases of conflict.  We take into account: -

·         Is the source of the information documented?   If owner1 has referenced (or better yet, supplied copies of) official documents such as marriage certificates while owner2 has no such source documentation, then we’ll believe owner1

·         Is the owner closely related to the subject?   We all take more interest in our close relatives, and have more and better information about them than we have about people who are only distantly and vaguely related.  On this criteria the best records should be those of dbarnes and robertb, both of whom are grandchildren of Hannah OLD. 

·         Does the record have good scrapbook entries?   While not actually proving anything, the presence of scrapbook entries gives us some comfort that the owner has had access to original sources of information about this person.  On this criteria the robertb record is probably the best.

·         What do we know about the record owner?  Some users have a reputation for carefully checking facts, while others simply copy whatever they get from the Internet.   Mirk562 has had a genealogy published about the OLD family, and so has done a lot of careful research that is well documented with sources and photos.  On this criteria the Mirk562 record should be the best.

 

If NZGDB calculated a field QualityIndex when records are inserted into the database, it would then be very easy to present the various OLD, Hannah(1860-1939) records in the right order.  But how do we calculate QualityIndex: -

 

·         Is the source of the information documented?  If sources are documented, then there will be a “fact detail” noting this.  Sometimes this will cite a registry entry, but more often this will simply record another GED file and the date of import: for example

o        Source        ThomasRobertson.FTW

o        ….Data

o        ……..Text   Date of Import: 4th September 2001

The program will not be able to distinguish between this and a source entry documenting an official BDM registry.

 

Perhaps the record has a copy (.JPG) of the actual certificate within the scrapbook.  The program can certainly detect that there are scrapbook entries of type “Certificate Copy”.   The presence of such records is strong evidence that sources have been well documented:  however such records are extremely rare.  

 

·         Is the owner closely related to the subject?  NZGDB is good at hiding living people, and you are encouraged to load your complete tree and let the system sort out privacy rather than spending time to strip out living people.  Many trees therefore include a record of the owner.  In such cases it would be quite easy to calculate a “relationship score”.  For example: -

o        0    if owner = subject.

o        1    One step removed: ie children, parents, spouse

o        2    Two steps removed.   Grandparents, grandchildren, and siblings

o        3    Three steps removed.  Ggrandparents, Ggrandchildren, Uncles and Aunts, Nephews and Nieces, Children-in-law

o        4    Four steps removed:  Gggrandparents, Gggrandchildren, 1st cousins, Parents of children-in-law.

And so on.  This is easy to define and calculate, provided that the owner’s record is within the tree. But is this relationship score a useful indicator of the quality of the record?

 

·         It would be very easy to calculate the number of scrapbook entries, although whether these are “good” entries can’t be determined.  Does this have any value?

·         I can’t think of any easy way of assessing the genealogy reputation of the record owner. The fact that Mirk562 has published a genealogy of the OLD family is NOT information that is recorded in the database.

 

So the only feasible idea that I can think of is to calculate the relationship score counting the distance between the subject and the record owner.  Is this worth doing?    Does anybody have a better idea?  Email me and let me know what you think.

 

Regards,

Robert Barnes,

NZGDB Developer